RC RANDOM CHAOS

OpenAI Pipeline Generates 948 Linux Privilege Escalation Exploits

· via The Hacker News

Original source

OpenAI AI Pipeline Generates 948 Linux Root Escalation Exploits in 30-Hour Run

The Hacker News →

OpenAI demonstrated an automated vulnerability pipeline that identified, validated, and produced exploits for Linux security flaws. A 30-hour run in mid-October 2026 analyzed 2,308 candidate kernel findings and reproduced 1,724 vulnerabilities, generating 948 working local privilege escalation (LPE) exploits and 291 information disclosure proofs-of-concept at a cost of $1.2 million. The pipeline stages include vulnerability discovery, dynamic validation in sandboxes, exploitability testing, patch generation, and human review before disclosure to maintainers.

The process revealed at least 10 novel Linux exploitation techniques, including new object-spraying and page grooming methods, data-only escalation approaches, and arbitrary type confusion tactics. This builds on 24 LPE exploits discovered by OpenAI’s GPT-5.5-Cyber model in June 2026.

The demonstration reflects a wider acceleration in vulnerability discovery driven by AI. Vulnerability disclosures increased from 5,045 in January 2026 to 10,740 in August 2026, with high-risk categories surging from 131 to 350 over the same period. The U.K. National Cyber Security Centre noted that defensive AI tools have compressed the zero-day-to-patch window but warned that poorly configured systems remain vulnerable. As AI capabilities become widely available to both defenders and attackers, comprehensive patching programs have become essential.

Read the full article

Continue reading at The Hacker News →

This is an AI-generated summary. Read the original for the full story.